Skip to content
GitHub

Engine Configuration

Engine configuration is found at Settings → Engine Configuration. For first-time setup or downloading an engine, see Engine Setup.

The Engine Configuration section shows a list of engines Masterboard knows about. Click an engine in the list to make it the active engine. To add one:

  • + Download engine… — download a common engine from within the app: Stockfish, Leela Chess Zero, or a human-style Maia network. See Engine Setup.
  • + Add local engine… — browse to a UCI engine binary already on your machine. Masterboard performs a UCI handshake to confirm it is a valid engine.

To remove an engine, use the × on its row; you can keep the downloaded files or delete them from disk. Use Refresh to re-scan the engines folder.

The hash (transposition table) is memory the engine reserves for caching positions it has already evaluated. Larger hash sizes allow deeper searches without redundant computation.

Recommended values:

System RAMHash setting
8 GB128–256 MB
16 GB256–512 MB
32 GB+512–1024 MB

Setting hash higher than about 75% of available RAM can cause swapping and slow the engine down. If you notice sluggishness, reduce this value.

The number of CPU threads the engine uses for analysis. More threads = stronger analysis, but uses more CPU, which may impact app responsiveness and other tasks.

A practical setting: total logical cores minus one. For example, on an 8-core CPU with 16 logical processors, set threads to 15.

To find your logical core count:

  • Windows: Task Manager → Performance → CPU → Logical processors
  • macOS: Terminal: sysctl -n hw.logicalcpu
  • Linux: Terminal: nproc

This sets the search depth used by automated game analysis when classifying every move in a game. The default is 22. Higher values produce more accurate classifications but take longer per game.

The number of candidate lines the engine shows is not set here — it is controlled directly in the Engine panel during live analysis (range 1–5). See Reading Results.

Hash and thread changes apply to both live analysis and full game analysis, and take effect the next time the engine starts a new search. A restart is not required.